Multimodal Data Fusion to Capture Dynamic Interactions between Built Environment and Vulnerable Older Adults
Houhao Liang, Azrin Jamaluddin, Kresimir Friganovic, Kirstie Neo, Raphael Han, Navrag Singh, Panos Mavros · 2026
Read Multimodal Data Fusion to Capture Dynamic Interactions between Built Environment and Vulnerable Older Adults by Houhao Liang, Azrin Jamaluddin, Kresimir Friganovic, Kirstie Neo, Raphael Han, Navrag Singh, Panos Mavros (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-11545.txt.
Introduction
Houhao Liang and six coauthors fuse data on how older adults move in cities. Their 2026 preprint studies micro-scale features of the built environment. Safe, inclusive mobility for vulnerable older adults is a rising priority. Urban planning needs to know what helps or hinders them.
Surveys and GIS audits give only a limited view today. They miss how small features shape real behavior and perception. A low curb or poor light can change a whole trip. The team fuses multimodal data to catch those dynamics. Dynamic interaction between people and places is the target. Better data can guide age-friendly street design.
Catalog record
| CREATOR | Houhao Liang, Azrin Jamaluddin, Kresimir Friganovic, Kirstie Neo, Raphael Han, Navrag Singh, Panos Mavros |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 20 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-11545
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Multimodal Data Fusion to Capture Dynamic Interactions between Built Environment and Vulnerable Older Adults Houhao Liang, Azrin Jamaluddin, Kresimir Friganovic, Kirstie Neo, Raphael Han, Navrag Singh, Panos Mavros Abstract Ensuring safe and inclusive mobility for vulnerable older adults is an emerging priority in urban planning. However, existing data sources such as surveys or GIS-based audits provide limited insight into how micro-scale built environment (BE) features influence real-world behavior and perception. This study presents a novel multimodal data-fusion approach that integrates wearable and environmental sensing to dynamically represent human-environment interactions and quantify the BE impacts on mobility among vulnerable older adults, specifically those with knee osteoarthritis or a history of falls. Data collected during naturalistic walking sessions in Singapore, are used to demonstrate this framework of synchronized streams from eye tracking, kinematic sensors, physiological monitors, GPS, and video recordings. Preliminary results show how AI-driven data fusion can uncover behaviorally and perceptually significant urban segments, providing a basis for actionable insights in inclusive design. This human-centered analytical approach advances the representation of urban environments from the perspective of vulnerable pedestrians, establishing a foundation for evidence-based, age-friendly city planning. Introduction As cities evolve toward human-centered and data-driven design, urban planners increasingly recognize the importance of understanding how built environment (BE) features shape everyday human experiences ( 14 ) . For vulnerable populations, particularly older adults with mobility impairments, such as those living with knee osteoarthritis (OA), or a history of falls, the BE can determine whether a space feels walkable, supportive, or hazardous. Designing inclusive and accessible BE is therefore essential to sustain mobility and promote social engagement among older adults. Achieving this vision requires accurate, data-driven representations of how individuals with mobility challenges dynamically interact with their surroundings in real-world contexts. Urban planning increasingly depends on the representation and quantification of urban environments to support evidence-based design and policy decisions. However, data used in current planning practices are often derived from surveys, interviews, and field audits, which provide subjective and aggregated perspectives. Such methods, while informative at the neighborhood scale, cannot capture the moment-to-moment behavioral, perceptual and physiological responses that occur during real-world walking. As a result, they provide limited understanding of how specific BE features influence physical adaptation, and walking stability, particularly among vulnerable older adults with mobility impairments. Motivated by recent advances in wearable sensing, computer vision, and robotics, new opportunities have emerged to represent and quantify urban experience through multimodal data fusion. Integrating visual, gaze, gait, and physiological signals, we examine how the BE dynamically influences human movement and perception at high temporal and spatial resolutions. Such evidence-based metrics can assist planners in evaluating comfort, safety, and accessibility at the micro scale, thereby complementing traditional environmental data. In this context, the present study introduces a multimodal data-fusion approach to dynamically represent and quantify BE and investigate the impacts on mobility and perception among vulnerable older adults. Through the combination of wearable and environmental sensing in real urban settings, the study provides an moment-to-moment representation of human–environment interaction that extends beyond static urban models. Figure 1: Overview of the multimodal data collection setup used to represent built-environment impacts on older adults’ mobility. Participants were equipped with eye-tracking glasses, wearable physiological sensors, and body-mounted IMUs while walking along urban routes in Singapore. A trailing GoPro camera provided third-person video. The integrated configuration captures synchronized gaze, gait, physiological, and environmental data for dynamic and fine-grained analysis of human–environment interactions. Related Work As populations age, older adults often experience declines in cognitive function and perceptual acuity that influence how they interact with their surrounding environment ( 12 ) . In this context, understanding and supporting these interactions have become growing priorities in both urban planning and public health ( 5 ) . Built environment (BE) features, including pavement conditions, pedestrian crossings, lighting, streetscape elements, and spatial layout, can substantially affect mobility behavior among older adults, either facilitating or constraining safe and confident movement in daily life ( 13 ; 22 ) . Given these impacts, it is essential to capture how BE features shape older adults’ mobility and behavior. Conventional data sources to support urban design such as interviews, surveys, and field audits have provided valuable insights into walkability and accessibility ( 3 ; 1 ) . However, these methods depend on self-reported experiences, which can be affected by recall bias, declining response rates, and the substantial time demands placed on participants and researchers. Spatial analysis techniques based on geographic data represent another category of assessment. These geospatial approaches utilize diverse datasets, such as environmental sensing (e.g. traffic noise) ( 21 ) , GPS logs ( 18 ) , and Street View Imagery ( 11 ) , to map and model BE features across large urban areas. Yet, these methods often rely on large-scale aggregated statistical approaches, and thus inadequately capture person-environment interactions at the micro-level, or account for individual differences. Hence, such assessments may overlook discrepancies in how environmental conditions affect different individuals ( 22 ) . More recently, individual-centered sensing has emerged as a promising method to capture how people interact with the BE in real-world settings, with the additional advantage of scalability for large population studies ( 8 ) . Participants can be equipped with wearable devices that continuously record real-time physiological (e.g., heart-rate variability (HRV), electrodermal activity (EDA)), kinematic (e.g., gait dynamics), and cognitive (e.g., gaze behavior) data. For example, motion sensors placed on the feet or waist have been used to quantify gait patterns and detect subtle changes in walking behavior in response to environmental challenges such as uneven pavement or slope ( 20 ; 17 ) . Similarly, wrist-worn sensors measuring HRV, EDA, and skin temperature effectively capture physiological stress responses to environmental stressors, such as crowded crossings or architectural obstacles ( 10 ) . Despite the increasing adoption of data-driven methods to investigate the BE’s impact, a critical gap remains in the integration of such multimodal data for micro-scale and holistic analysis. Few studies have comprehensively combined visual, gaze, gait, and spatial information to understand how older adults with mobility impairments interact with detailed BE features during everyday walking. Addressing this gap, the present study aims to develop a data-driven representation of BE features and quantify their micro-scale impacts on mobility and behavior among vulnerable older adults. Leveraging advances in computer vision, robotics, and spatial mapping, this study enables dynamic and fine-grained representation of BE features in real urban contexts. Study Overview The study focuses on community-dwelling older adults aged 45 to 90 years living in the Yio Chu Kang district of Singapore. This area represents a dense residential environment typical of mature Singapore neighborhoods. The selected walking routes include covered walkways, pedestrian crossings, green corridors, and commercial access paths, providing a realistic testbed for examining how built environment (BE) features influence movement and perception in older adults. Participants were purposively recruited through healthcare institutions and community networks. Eligibility required either a clinical diagnosis of knee osteoarthritis (OA) or at least one fall within the past two years, both of which are associated with reduced balance confidence and altered gait. The participants thus represent a vulnerable population segment that frequently encounters physical and cognitive challenges during daily mobility. A naturalistic observation protocol is designed to capture